Align methods for #858

[mortonjt]

This pull request adds alignment functionality, to allow the biom table to be aligned against metadata (represented as pd.DataFrame) and trees (represented as skbio.TreeNode).

I have made this a little more general, allowing these alignment methods to be performed on both axes (sample and observations).

Once we have agreed on the basic user interface, I will push in the remaining unittests.

Outstanding questions

1. Should we enable an inplace option? Maybe useful for massive datasets.
2. Should we enable a how or join flag, specifying if this is going to be an inner, outer, left or right join? Right now, I'm only implementing an inner flag since those are my only use cases. But if there is a compelling use case, I am open for discussion.

[wasade]

Thanks! There are a few suggestions to make the code more direct. If this is based on an existing implementation, it may be worth verifying the sorting works as expected as the prior use of sort_f I don't believe would produce a correct table

[wasade]

Inner is fine. Tests would be good :) Possible to address the failure items (looks like lint), and to also include a changelog note?

[mortonjt]

@wasade thank you for the thorough review - still a little struck how the tree alignment could have been off this whole time.
The code here seems to be working, but I may need to double check my other workflows once this is merged in ...

EDIT : there are some failing tests that don't appear to be relevant to this PR, for instance.

test_table.py:668: in test_from_hdf5_custom_parsers
    self.assertIn(m['BODY_SITE'], ('GUT', 'SKIN'))
E   AssertionError: b'GUT' not found in ('GUT', 'SKIN')

[wasade]

Thanks! As written, the skbio tests would never have ran btw. Can CI be updated as well to install skbio? I would like CI to pass for at least some of the builds in the matrix before merge, and I can adjust later some builds fail

[mortonjt]

Just added in doctests - I believe the outstanding comments have been addressed.

[wasade]

Wonderful, thank you @mortonjt!!

[mortonjt]

Thank you!

…

On Wed, Jul 28, 2021 at 10:42 AM Daniel McDonald ***@***.***> wrote: Merged #859 <#859> into master. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#859 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXLPNLGDQFSF6RFCOETT2AXQNANCNFSM5BBBK7MQ> .